动态因子模型在我国人均GDP对肺结核发病率影响研究中的应用  

Application of Dynamic Factor Model in the Research of Per Capita Gross Domestic Production on the Incidence of Tuberculosis in China

作　　者：杨丽娟 段禹[1] 张燕杰 操龙挺 潘贵霞[1] 叶冬青[1,2] 王静 Yang Lijuan;Duan Yu;Zhang Yanjie(Department of Epidemiology and Biostatistics,School of Public Health,Anhui Medical University(230032),Hefei)

机构地区：[1]安徽医科大学流行病与卫生统计学系,230032 [2]安徽省重大自身免疫性疾病重点实验室

出　　处：《中国卫生统计》2019年第3期351-353,357,共4页Chinese Journal of Health Statistics

基　　金：安徽省质量工程教学研究重大项目(2016jyxm0378);安徽省教育厅自然科学研究重点项目(KJ2017A164);安徽省教育厅高校优秀青年骨干人才国外访学研修项目(gxfx2017008)

摘　　要：目的用动态因子模型提取并比较我国各地区肺结核发病率共同趋势,分析其与人均国内生产总值(GDP)的关系。方法用Stata 14.0基于动态因子模型分别对我国2004-2015年各地区肺结核发病率标准化值提取共同因子。用SPSS 23.0做滞后相关分析。以共同因子为因变量,人均月度GDP(千元)相应滞后期为自变量做线性回归分析。结果我国肺结核年均发病率(1/10万)最高的三个省是新疆(178.65)、贵州(146.60)和海南(116.58)。各地区肺结核发病率共同因子均在下降,中部、东北和西部北方地区下降较快。滞后相关分析显示中部和东部南方地区的人均GDP对发病率共同因子的影响滞后二期,东部北方滞后三期,西部南方、西部北方和东北地区滞后一期。发病率共同因子与人均GDP负相关,人均月度GDP每增加1千元,东部南方地区肺结核发病率(1/10万)减少34.4%,北方地区减少32.5%;西部南方地区减少14.9%,北方地区减少75.2%;中部地区减少68.8%,东北地区减少45.3%。结论我国肺结核发病趋势在降低,且有一定地区差异。人均GDP对肺结核发病率的影响滞后1-3个月。应加大西部南方地区对肺结核防治的经济投入。动态因子模型可应用于肺结核发病数据的分析。Objective We aimed to extract and observe the common trend of tuberculosis incidence in different regions of China based on dynamic factor model,and to analyze the relationship between tuberculosis incidence and economic factors.Methods Based on dynamic factor model,we extracted the dynamic factors of standardized tuberculosis incidence in different regions in 2004-2015 with Stata 14.0 software.The lag correlation analysis was performed with the software of SPSS 23.0.Linear regression analysis was conducted with dependent variable of dynamic factor and independent variable of the monthly per capita GDP with corresponding lag in different regions.Results Xinjiang(178.65),Guizhou(146.60)and Hainan(116.58)were the three provinces with the highest annual incidence(/100000)of tuberculosis in China.Dynamic factors of tuberculosis incidence in different regions are on a decline.Incidence in the central and northeast regions fall the fastest.The GDP per capital lagged two months in the central region and south area of eastern region,three months in the north area of eastern region,and one month in the south and north areas of western region as well as the north-eastern region according to the lag correlation analysis.Dynamic factors were negatively correlated with monthly per capita GDP.Tuberculosis incidence(/100000)had a decrease of 34.4% and 32.5% in the south and north areas of eastern region,14.9% and 75.2% in the south and north areas of western region,68.8% and 34.4% in central and northeast regions for every one thousand yuan increase in regional monthly per capita GDP,respectively.Conclusion Incidence trend of tuberculosis is on a decline in our country,and it varied in different regions.The impact of GDP per capita on tuberculosis incidence lagged 1 to 3 months.Economic investment in controlling tuberculosis should be increased in the south areas of western region.Dynamic factor model could be applied to the analysis of tuberculosis incidence data.

关 键 词：肺结核 动态因子模型 时间序列 线性回归 

分 类 号：O211.61[理学—概率论与数理统计] R521[理学—数学]